ABSTRACT
Rationale Many patient with end-stage fibrotic lung disease die in the hospital due to dependence on High Flow Nasal Cannula (HFNC), not available at home. Despite available technical maturity, HFNC is not sufficiently available due to size and portability, energy requirement and oxygen (O2) capacity. This has become particularly problematic during the current Covid-19 pandemic, when many had to die alone in hospital due to these constraints. ObjectiveThe objective of this investigation was to appreciate the number of patients who die in hospital on comfort care, who could benefit from HFNC at home to support person-centered care (self-determination, family involvement, comforts of home) at end of life (EOL). MethodsWe collected data from a convenience sampling of patients who received care at a 449-bed acute-care community hospital in Southern California between June 2020 and June 2021. Data from this retrospective review of the electronic health record included demographics (age, gender, BMI and ethnicity). Other variables collected were length of stay (LOS), Covid status (+/-), comfort care orders, and HFNC requirement. Data were analyzed for frequencies, means and percentages. Chi square (categorical) and t- tests (continuous) were performed to determine statistical significance and Pearson r (categorical) and eta (continuous) were performed to test strength of association.ResultsOf the sample (n = 91), mean age was 78 years (+/- 10.6) and mostly female (38.8%, n = 42). Mean LOS was 13.7 days (+/- 12.1). Most (71.4%) patients in the sample were Hispanic (n = 65). 63 patients had orders for comfort care (69.2%), and 61 patients were Covid positive (67%). There was a statistically significant difference in mean flow rate (p = 0.022, η = 0.564) and fi02 (p < 0.001, η = 0.688) for patients discharged to hospice vs. those who died in hospital. For patients who died in hospital, mean fi02 was 0.94 (12.2) and mean flow rate (in liters) was 48.6 (16.4). ConclusionsThe pandemic has highlighted many healthcare disparities in the United States, and made apparent the needs of persons with fibrotic lung disease at EOL. This investigation revealed that most patients in hospital opting for comfort care died in hospital as their needs for increased flow rates and fi02 far exceeds what is currently available for outpatient use of HFNC. Investments should be made into developing technologies to support these individuals with the benefits of decreased need for hospitalization and promoting self-determination at EOL.
ABSTRACT
Rationale Barotrauma has been observed in ARDS patients with COVID-19 at greater incidence than ARDS not related to COVID-19, despite low-volume ventilation. We analyzed data of intubated patients with confirmed COVID-19 and bilateral opacities, to determine risk factors for barotrauma, with consideration to demographic information, pre-existing conditions, certain laboratory findings, therapies and ventilation parameters. Methods Data for this pilot study were collected from a convenience sampling of patients at a multicenter healthcare system in southern California between March and May 2020. Cases of Barotrauma were identified by intensivists providing care, and a denominator was selected from a larger database. Data from this retrospective review of the electronic health record included demographics (age, gender, height, weight and ethnicity), certain medications and therapies, select serum lab values, ventilation parameters and survival. Cases were grouped as Barotrauma vs. no Barotrauma. T tests and chi square were used as tests of difference, Pearson r and eta were used as tests of association. Results Twenty-two cases had complete data available for analysis. Mean age was 62(+/-15.8), 31.8% female, and mean P/f ratio was 112.3(+/-42.6). Of the total sample, 13 (59.1%) patients expired. Thirteen cases (59.1%) suffered barotrauma, and 9(40.1%) did not. Serum ferritin (p = 0.046, η = 1.0) and administration of convalescent plasma (p = 0.011, r = 0.555), were statistically significant and highly correlated in the barotrauma group. Certain ventilation parameters were also statistically significant and highly correlated with barotrauma including fiO2 (p = 0.027, η = 0.759), PIP (p = 0.004, η = 0.855), Pplat (p = 0.002, η = 0.835), Pdrive (p = 0.003, η = 0.772), Cstat (p = 0.044, η = 0.893) and Pmean (p = 0.029, η = 0.804). Of the 13 cases included with Barotrauma, 9 (69.2%) had pneumothorax, 8 (61.5%) had pneumomediastinum, 7 (53.8%) had subcutaneous emphysema, 1 (7.7%) had pneumopericardium and 1 (7.7%) had pneumoperitoneum. Eight cases (61.5%) had multiple subtypes of barotrauma. Mean VT (mL/kg) was 6.9mL and median days to onset of barotrauma after intubation was 2 days. Conclusions Our analysis supports other findings associating ventilation parameters with the incidence of barotrauma in COVID-19 related illness. Although there is literature available to link elevated serum ferritin to COVID-19 related illness, its relationship to barotrauma has not been established. Convalescent plasma administration is not otherwise associated with barotrauma in the current literature. These findings should be confirmed by a larger, well powered investigation.
ABSTRACT
Rationale Predictive models of mortality for COVID 19 related illness have yet to be sufficiently tested. We analyzed the data of confirmed COVID-19 cases to determine possible predictors of mortality in hospitalized patients with COVID-19 related illness. Consideration was given to demographic information, pre-existing patient conditions and acute management. Methods We collected data from a convenience sampling of patients who received care at a multicenter healthcare system in southern California between March and August 2020. Data from this retrospective review of the electronic health record included demographics (age, gender, BMI and ethnicity), comorbidities (HTN, DM, CHF, chronic lung disease), lab values, certain medications (statins, antihypertensives, experimental therapies) and mortality risk by way of the updated Charlson Comorbidity index (CCI), which includes 12 categories of comorbidity predicting 1-year mortality (Quan et al., 2011). Cases were grouped as expired vs. survived. Chi square (categorical) and t-tests (continuous) were performed to determine statistical significance and Pearson r (categorical) and eta (continuous) were performed to test strength of association. Logistic regression was performed to test a predictability model for variables with the highest degree of statistical significance and correlation with mortality. Results Of 5,559 cases reviewed, 139 cases had complete data available for analysis. Of the total sample, mean age was 61(+/-16.1) and 38.8% female. Thirtyseven (26.6%) patients expired and 102 (73.4%) patients survived. Four variables were found to have a statistically significant, yet mild association with mortality. These four variables included: Betablocker use (p = 0.028, r = 0.186), CHF (p = 0.018, r = 0.200), chronic lung disease (p = 0.011, r = 0.217) and CCI (p 0.03, r = 0.216). Two variables had statistically significant and strong correlations to mortality;Age (p = 0.004, r = 0.636) and CRP (p = 0.003, r = 0.937). CRP and age accounted for 73.4% predictive value of mortality in our model. Conclusions Our analysis contributes further understanding of predictors of mortality in hospitalized patients with a confirmed diagnosis of COVID-19. HTN and DM were not associated with increased risk of mortality. Use of beta blockers, CHF and chronic lung disease had mild degree of association with mortality. Age and CRP served as strong predictors of mortality in our model. This retrospective convenience sample study was not conducted by way of a powered investigation. Findings should be confirmed by way of a well powered investigation before inferring utility of the model.
ABSTRACT
RationaleMuch remains unknown about individual risk factors for COVID 19 related illness. We analyzed the data of confirmed COVID-19 cases to determine individual risk factors for COVID-19 related illness, with consideration to demographic information, pre-existing patient conditions and their relationship to severity of COVID-19 related illness.MethodsWe collected data from a convenience sampling of patients who received care at a multicenter healthcare system in southern California between March and August 2020. Data from this retrospective review of the electronic health record included demographics (age, gender, BMI and ethnicity), blood type, certain medications and mortality risk by way of the updated Charlson Comorbidity index (CCI), which includes 12 categories of comorbidity predicting 1-year mortality (Quan et al., 2011). Cases were stratified by severity (discharged home from Emergency Department, admitted as an inpatient to a non-acute care setting, admitted as a patient to an acute care setting, need for intubation, and survivability. ANOVA was used as a test of difference;Pearson r and eta were used as tests of association. ResultsOf 5,559 cases reviewed, 391 cases had complete data available for analysis. Eighteen cases (0.05%) were discharged home from the Emergency Department, 65 (16.6%) were admitted to a non-acute setting and 308 (78.8%) were admitted to an acute care setting. 176 (45%) cases required endotracheal intubation, and of the total sample, 100 (25.6) patients expired. Age (p < 0.001, r = 0.666), gender (p < 0.001, η = 0.351), BMI (p = 0.042, η = 0.906) and history of Diabetes (p < 0.001, r = 0.309) were strongly correlated with patient disposition (discharge vs. non-acute, vs. acute). Age (p <0.001, r = 0.557), Gender (p <0.001, r = 0.307), BMI (p = 0.021, η = 0.915), and Beta Blocker use (p <0.001, r = 0.306) were strongly correlated to risk for intubation, and Age (p < 0.001, η = 605), BMI (p = 0.015, η = 0.905) and CCI (p < 0.001, η = 0.3), were strongly correlated with survivability. ConclusionsOur analysis contributes further understanding of risk stratification of severity of illness in patients with a confirmed diagnosis of COVID-19. Hypertension has been frequently reported in the literature to be an important risk factor for COVID-19 related illness. However, in our analysis, it was not strongly correlated to the outcomes under investigation.